public inbox for linux-media@vger.kernel.org
 help / color / mirror / Atom feed
From: Hans Verkuil <hverkuil@xs4all.nl>
To: Mauro Carvalho Chehab <m.chehab@samsung.com>
Cc: linux-media@vger.kernel.org, laurent.pinchart@ideasonboard.com,
	s.nawrocki@samsung.com, ismael.luceno@corp.bluecherry.net,
	pete@sensoray.com, sakari.ailus@iki.fi,
	Hans Verkuil <hans.verkuil@cisco.com>
Subject: Re: [REVIEWv3 PATCH 04/35] videodev2.h: add initial support for complex controls.
Date: Tue, 11 Mar 2014 21:23:26 +0100	[thread overview]
Message-ID: <531F70BE.5060307@xs4all.nl> (raw)
In-Reply-To: <20140311163414.0c1f788e@samsung.com>

Hi Mauro,

On 03/11/2014 08:34 PM, Mauro Carvalho Chehab wrote:
> Em Mon, 17 Feb 2014 10:57:19 +0100
> Hans Verkuil <hverkuil@xs4all.nl> escreveu:
> 
>> From: Hans Verkuil <hans.verkuil@cisco.com>
>>
>> Complex controls are controls that can be used for compound and array
>> types. This allows for more complex data structures to be used with the
>> control framework.
>>
>> Such controls always have the V4L2_CTRL_FLAG_HIDDEN flag set. Note that
>> 'simple' controls can also set that flag.
>>
>> The existing V4L2_CTRL_FLAG_NEXT_CTRL flag will only enumerate controls
>> that do not have the HIDDEN flag, so a new V4L2_CTRL_FLAG_NEXT_HIDDEN flag
>> is added to enumerate hidden controls. Set both flags to enumerate any
>> controls (hidden or not).
>>
>> Complex control types will start at V4L2_CTRL_COMPLEX_TYPES. In addition, any
>> control that uses the new 'p' field or the existing 'string' field will have
>> flag V4L2_CTRL_FLAG_IS_PTR set.
>>
>> While not strictly necessary, adding that flag makes life for applications
>> a lot simpler. If the flag is not set, then the control value is set
>> through the value or value64 fields of struct v4l2_ext_control, otherwise
>> a pointer points to the value.
>>
>> Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
>> Reviewed-by: Sylwester Nawrocki <s.nawrocki@samsung.com>
>> ---
>>  include/uapi/linux/videodev2.h | 11 +++++++++--
>>  1 file changed, 9 insertions(+), 2 deletions(-)
>>
>> diff --git a/include/uapi/linux/videodev2.h b/include/uapi/linux/videodev2.h
>> index 6ae7bbe..4d7782a 100644
>> --- a/include/uapi/linux/videodev2.h
>> +++ b/include/uapi/linux/videodev2.h
>> @@ -1228,6 +1228,7 @@ struct v4l2_ext_control {
>>  		__s32 value;
>>  		__s64 value64;
>>  		char *string;
>> +		void *p;
> 
> Hmm... don't we have already "string" for pointers? Also, calling it
> as "p" inside an userspace api doesn't seem to nice ("ptr" would be
> better).
> 
> Btw, you likely already noticed the mess, as, when you added 
> this email's comment, you said that complex controls could
> either use "p" or "string".
> 
> Nack. It should just use "string". Let's not add even more complexity
> to this "complex" controls.

It is really, really weird to refer to a matrix type or struct type
through a 'string'. Also, the type of 'string' is a char pointer
which is totally wrong unless the type is really a string.

Finally there is a slight difference in handling strings: the size of
the memory 'string' points to just has to be large enough to store the
0-terminated string instead of being able to store the maximum possible
size of the string (max + 1). This is IMHO a design mistake regarding
strings on my part. I should always have required that the memory is
always sizes for the worst-case.

I might take another look if I can change this behavior without
breaking existing applications.

I don't mind changing 'p' to 'ptr', no problem.

> 
>>  	};
>>  } __attribute__ ((packed));
>>  
>> @@ -1252,7 +1253,10 @@ enum v4l2_ctrl_type {
>>  	V4L2_CTRL_TYPE_CTRL_CLASS    = 6,
>>  	V4L2_CTRL_TYPE_STRING        = 7,
>>  	V4L2_CTRL_TYPE_BITMASK       = 8,
>> -	V4L2_CTRL_TYPE_INTEGER_MENU = 9,
>> +	V4L2_CTRL_TYPE_INTEGER_MENU  = 9,
>> +
>> +	/* Complex types are >= 0x0100 */
>> +	V4L2_CTRL_COMPLEX_TYPES	     = 0x0100,
>>  };
> 
> Not sure if I got why you're calling it as "TYPES" and saying that
> everything >= 0x100 is complex. What's your idea here?

Types divide into 'simple' types (the ones we have today) and the
new complex (compound?) types. I need an easy way to determine
which it is, so any types with values >= 0x100 fall into the latter
category. Later in the patch series when such types are actually
added the meaning becomes more obvious.
 
> Also, at least for me with my engineering formation, "complex"
> means a number with an imaginary component.
> 
> And yes, we do have complex numbers on some usecases for
> V4L (for example, SDR in-phase/quadrature, e. g.  I/Q
> representation can be seen as an array of complex numbers).
> 
> So, I won't doubt that someone might propose some day to add a
> way to set a complex number via a V4L2 control.
> 
> So, please use a better naming here to avoid troubles.

COMPOUND_TYPES? Not a bad name, I think.

Regards,

	Hans

> 
>>  /*  Used in the VIDIOC_QUERYCTRL ioctl for querying controls */
>> @@ -1288,9 +1292,12 @@ struct v4l2_querymenu {
>>  #define V4L2_CTRL_FLAG_SLIDER 		0x0020
>>  #define V4L2_CTRL_FLAG_WRITE_ONLY 	0x0040
>>  #define V4L2_CTRL_FLAG_VOLATILE		0x0080
>> +#define V4L2_CTRL_FLAG_HIDDEN		0x0100
>> +#define V4L2_CTRL_FLAG_IS_PTR		0x0200
>>  
>> -/*  Query flag, to be ORed with the control ID */
>> +/*  Query flags, to be ORed with the control ID */
>>  #define V4L2_CTRL_FLAG_NEXT_CTRL	0x80000000
>> +#define V4L2_CTRL_FLAG_NEXT_HIDDEN	0x40000000
>>  
>>  /*  User-class control IDs defined by V4L2 */
>>  #define V4L2_CID_MAX_CTRLS		1024
> 
> 


  reply	other threads:[~2014-03-11 20:23 UTC|newest]

Thread overview: 66+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2014-02-17  9:57 [REVIEWv3 PATCH 00/35] Add support for complex controls, use in solo/go7007 Hans Verkuil
2014-02-17  9:57 ` [REVIEWv3 PATCH 01/35] v4l2-ctrls: increase internal min/max/step/def to 64 bit Hans Verkuil
2014-02-17  9:57 ` [REVIEWv3 PATCH 02/35] v4l2-ctrls: add unit string Hans Verkuil
2014-02-17  9:57 ` [REVIEWv3 PATCH 03/35] v4l2-ctrls: use pr_info/cont instead of printk Hans Verkuil
2014-02-17  9:57 ` [REVIEWv3 PATCH 04/35] videodev2.h: add initial support for complex controls Hans Verkuil
2014-03-11 19:34   ` Mauro Carvalho Chehab
2014-03-11 20:23     ` Hans Verkuil [this message]
2014-03-11 23:48       ` Mauro Carvalho Chehab
2014-02-17  9:57 ` [REVIEWv3 PATCH 05/35] videodev2.h: add struct v4l2_query_ext_ctrl and VIDIOC_QUERY_EXT_CTRL Hans Verkuil
2014-03-11 19:42   ` Mauro Carvalho Chehab
2014-03-11 20:29     ` Hans Verkuil
2014-03-11 23:35       ` Mauro Carvalho Chehab
2014-02-17  9:57 ` [REVIEWv3 PATCH 06/35] v4l2-ctrls: add support for complex types Hans Verkuil
2014-03-11 20:14   ` Mauro Carvalho Chehab
2014-03-11 20:43     ` Hans Verkuil
2014-03-11 23:43       ` Mauro Carvalho Chehab
2014-02-17  9:57 ` [REVIEWv3 PATCH 07/35] v4l2: integrate support for VIDIOC_QUERY_EXT_CTRL Hans Verkuil
2014-02-17  9:57 ` [REVIEWv3 PATCH 08/35] v4l2-ctrls: create type_ops Hans Verkuil
2014-03-11 20:22   ` Mauro Carvalho Chehab
2014-03-11 20:49     ` Hans Verkuil
2014-03-11 23:56       ` Mauro Carvalho Chehab
2014-02-17  9:57 ` [REVIEWv3 PATCH 09/35] v4l2-ctrls: rewrite copy routines to operate on union v4l2_ctrl_ptr Hans Verkuil
2014-02-17  9:57 ` [REVIEWv3 PATCH 10/35] v4l2-ctrls: compare values only once Hans Verkuil
2014-02-17  9:57 ` [REVIEWv3 PATCH 11/35] v4l2-ctrls: prepare for matrix support: add cols & rows fields Hans Verkuil
2014-03-12 10:34   ` Mauro Carvalho Chehab
2014-02-17  9:57 ` [REVIEWv3 PATCH 12/35] v4l2-ctrls: replace cur by a union v4l2_ctrl_ptr Hans Verkuil
2014-03-12 10:38   ` Mauro Carvalho Chehab
2014-02-17  9:57 ` [REVIEWv3 PATCH 13/35] v4l2-ctrls: use 'new' to access pointer controls Hans Verkuil
2014-03-12 10:40   ` Mauro Carvalho Chehab
2014-02-17  9:57 ` [REVIEWv3 PATCH 14/35] v4l2-ctrls: prepare for matrix support Hans Verkuil
2014-03-12 10:42   ` Mauro Carvalho Chehab
2014-03-12 12:21     ` Hans Verkuil
2014-03-12 13:00       ` Mauro Carvalho Chehab
2014-03-12 13:00       ` Mauro Carvalho Chehab
2014-03-12 13:41         ` Hans Verkuil
2014-03-12 13:44         ` Sylwester Nawrocki
2014-02-17  9:57 ` [REVIEWv3 PATCH 15/35] v4l2-ctrls: type_ops can handle matrix elements Hans Verkuil
2014-02-17  9:57 ` [REVIEWv3 PATCH 16/35] v4l2-ctrls: add matrix support Hans Verkuil
2014-02-17  9:57 ` [REVIEWv3 PATCH 17/35] v4l2-ctrls: return elem_size instead of strlen Hans Verkuil
2014-02-17  9:57 ` [REVIEWv3 PATCH 18/35] v4l2-ctrl: fix error return of copy_to/from_user Hans Verkuil
2014-02-17  9:57 ` [REVIEWv3 PATCH 19/35] DocBook media: document VIDIOC_QUERY_EXT_CTRL Hans Verkuil
2014-03-12 14:13   ` Mauro Carvalho Chehab
2014-03-13  7:58     ` Hans Verkuil
2014-02-17  9:57 ` [REVIEWv3 PATCH 20/35] DocBook media: update VIDIOC_G/S/TRY_EXT_CTRLS Hans Verkuil
2014-03-12 14:20   ` Mauro Carvalho Chehab
2014-03-13 12:18     ` Hans Verkuil
2014-02-17  9:57 ` [REVIEWv3 PATCH 21/35] DocBook media: fix coding style in the control example code Hans Verkuil
2014-02-17  9:57 ` [REVIEWv3 PATCH 22/35] DocBook media: update control section Hans Verkuil
2014-02-19 23:15   ` Sakari Ailus
2014-03-12 14:27   ` Mauro Carvalho Chehab
2014-02-17  9:57 ` [REVIEWv3 PATCH 23/35] v4l2-controls.txt: update to the new way of accessing controls Hans Verkuil
2014-02-17  9:57 ` [REVIEWv3 PATCH 24/35] v4l2-ctrls/videodev2.h: add u8 and u16 types Hans Verkuil
2014-03-12 14:44   ` Mauro Carvalho Chehab
2014-02-17  9:57 ` [REVIEWv3 PATCH 25/35] DocBook media: document new u8 and u16 control types Hans Verkuil
2014-02-17  9:57 ` [REVIEWv3 PATCH 26/35] v4l2-ctrls: fix comments Hans Verkuil
2014-02-17  9:57 ` [REVIEWv3 PATCH 27/35] v4l2-ctrls/v4l2-controls.h: add MD controls Hans Verkuil
2014-02-17  9:57 ` [REVIEWv3 PATCH 28/35] DocBook media: document new motion detection controls Hans Verkuil
2014-02-17  9:57 ` [REVIEWv3 PATCH 29/35] v4l2: add a motion detection event Hans Verkuil
2014-02-17  9:57 ` [REVIEWv3 PATCH 30/35] DocBook: document new v4l " Hans Verkuil
2014-02-17  9:57 ` [REVIEWv3 PATCH 31/35] solo6x10: implement the new motion detection controls Hans Verkuil
2014-02-17  9:57 ` [REVIEWv3 PATCH 32/35] solo6x10: implement the motion detection event Hans Verkuil
2014-02-17  9:57 ` [REVIEWv3 PATCH 33/35] solo6x10: fix 'dma from stack' warning Hans Verkuil
2014-02-17  9:57 ` [REVIEWv3 PATCH 34/35] solo6x10: check dma_map_sg() return value Hans Verkuil
2014-02-17  9:57 ` [REVIEWv3 PATCH 35/35] go7007: add motion detection support Hans Verkuil
2014-02-19  8:28 ` [REVIEWv3 PATCH 00/35] Add support for complex controls, use in solo/go7007 Ricardo Ribalda Delgado
2014-02-19  8:54   ` Hans Verkuil

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=531F70BE.5060307@xs4all.nl \
    --to=hverkuil@xs4all.nl \
    --cc=hans.verkuil@cisco.com \
    --cc=ismael.luceno@corp.bluecherry.net \
    --cc=laurent.pinchart@ideasonboard.com \
    --cc=linux-media@vger.kernel.org \
    --cc=m.chehab@samsung.com \
    --cc=pete@sensoray.com \
    --cc=s.nawrocki@samsung.com \
    --cc=sakari.ailus@iki.fi \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox